Builder Design Pattern and Cyclomatic Complexity Reduction

In this post, I will explain in detail how I used the builder design pattern together with the bit mask field technique to reduce complexity. As a result, I decreased the cyclomatic complexity from 24 to 0 and reduced development time from hours to minutes.

The rest of this post is organized as follows sections:

The problem explains the problem I was solving.
Creating the bit mask field presents a solution to the problem.
Cyclomatic Complexity discusses cyclomatic complexity and calculates this metric for the solution presented in the section above.
[Builder…

The rest of this post is organized as follows sections:

The problem explains the problem I was solving.
Creating the bit mask field presents a solution to the problem.
Cyclomatic Complexity discusses cyclomatic complexity and calculates this metric for the solution presented in the section above.
Builder design pattern explains the builder design pattern and describes the refactoring process that removed complexity.
Conclusion summarizes this post.
References provides references related to the topics discussed here.

All the examples shown here are in C++.

The problem

I’m working in an extreme scenario where every millisecond and byte counts. In this type of scenario, we need to design our solutions with extra care regarding the details. For example, how can we represent 32 Boolean variables for storage and transmission over computer networks?

In C++, the simplest way is to create 32 different Boolean variables. A Boolean variable in C++ occupies 1 byte (8 bits). So, we will need 32 × 1 = 32 bytes to represent all of these variables. This is not so bad. However, these variables are part of an object, and we could have one billion of these objects. Then, we would consume 32 gigabytes (GB) only to represent these Boolean variables.

We can improve this by using a technique called a bit mask field, where each bit of an integer variable is used to represent a Boolean variable. Considering an integer variable of 4 bytes (32 bits), we need only one integer variable to represent all 32 Boolean variables. With one billion objects, we would now consume 4 GB instead of 32 GB. This is an 87.5% savings.

This is a huge saving, considering that we need to store these variables on disk and transport them via a network between many different systems. The savings here are in both money and time. For the end user, this means they can get results more quickly.

You may be wondering, why don’t we use this technique in all systems to represent our set of Booleans? The reason is pretty simple: we are reducing simplicity to cut costs. In a non-extreme scenario, using this technique is over-engineering. Instead of saving money, you are creating complexity for your team to handle.

When we decide to use this technique, we need to resolve the problem: how can we create an integer variable to represent Boolean variables in our codebase in an easy-to-maintain way? The short answer is to encapsulate the complexity, which is explained in the rest of this post.

Creating the bit mask field

For a bit mask field, each bit of an integer variable represents a field. We need a way to set and check if a field is true in this variable. For both operations, we need to use bitwise operators. For example, let’s say we have the int variable fields, and we need to set and check the value of field 3 (bit 2, since we start counting from 0). To set field 3 to true in fields, we can do:

fields |= (1 << 2);

The << is the left shift operator that shifts the number by 2 positions to the left. So, this will create the number 4 (100 in binary).

To check if field 3 is true in fields, we can do:

int mask = (1 << 2);
(fields & mask) == mask;

The & is the bitwise AND operator that returns 1 when both corresponding bits in fields and mask are 1, and 0 otherwise. In other words, field 3 is true only when the bitwise AND operator returns mask as the result.

These are the operations in isolation; our solution needs to create the variable fields and set each field based on some conditions, as shown below:

bool condition1(){
//call complexity methods here
return false; // just as an example
};

bool condition2(){
//call complexity methods here
return true; // just as an example
};

int createFields(){
int fields = 0;

if(condition1()){
fields |= (1 << 1);
}

if(condition2()){
fields |= (1 << 2);
}

return fields;
}

Cyclomatic complexity

The cyclomatic complexity metric was developed by Thomas J. McCabe in the paper “A Complexity Measure”. Basically, it is a metric used to calculate how many paths your code has. To calculate this metric in your codebase, you can use lizard, which is an extensible Cyclomatic Complexity Analyzer for many programming languages.

The function createFields, presented in the last section, has a cyclomatic complexity of 3: 1 for the function itself, +1 for the first condition, and +1 for the second condition.

One question naturally emerges: What is a good cyclomatic complexity? The answer is not simple. There is no magic number, and this metric is just an indicator of potential issues in your codebase. Generally, smaller numbers are better.

In my opinion, the most important aspect is not just the number or the conditions in the code. The biggest problems are complexity and dependencies, which this metric can help you discover. Of course, you do not need this metric to identify complexity; you can do it yourself by reading the code. However, in a large codebase, this metric helps you decide which places to review first.

Let’s go back and analyze the function createFields. The cyclomatic complexity metric counts +1 for each condition, which is only an indicator. For each of these conditions, the function createFields is coupled with the functions condition1 and condition2, and the path our program will take depends on the result of these functions.

In other words, the client of this code—the caller—is responsible for knowing all these details and handling them. The most basic client is the tests. In this case, to test if fields1 is set correctly, we need to either set up everything that the functions condition1 and condition2 need or mock these functions. Even though I want to check just field1, I need to set up all the dependencies of createFields, not just the dependencies for field1.

We can reduce the complexity of the function createFields by using the Builder design pattern, as explained in the next section.

Builder design pattern

The Builder design pattern is a creational patterns that helps encapsulate the complexity of creating an object. It is described in the famous book Design Patterns: Elements of Reusable Object-Oriented Software by Erich Gamma, Richard Helm, Ralph Johnson, and John Vlissides, also known as the Gang of Four (GoF). This book is a classic and a must-read for experienced software engineers. I agree with Martin Fowler - it’s a great book, but not an easy read. It’s best approached when you’re comfortable with object-oriented design. As Martin Fowler noted in 2004, “In my view, the Gang of Four is the best book ever written on object-oriented design”.

The function createFields is responsible for creating the variable fields, which maps many Boolean variables into a single integer variable to save memory. This is a very good candidate to be simplified using the Builder design pattern. So, we can rewrite createFields using the Builder design pattern to encapsulate and reduce complexity. Below, we present the new code. You can notice some differences here. Now, we have classes instead of just a function. The BitMaskField class encapsulates the integer variable that represents the Boolean variables and provides a clear API for clients. The BitMaskFieldBuilder class is used to construct BitMaskField objects. The withField1 and withField2 methods handle the creation of each field, respectively.

Besides the improved organization, one major benefit is that the code is now decoupled. To use field1, callers do not need to know anything about the dependencies of other fields. All the complexity of creating an integer variable to map many Boolean variables is encapsulated in the BitMaskFieldBuilder class, and each field is completely decoupled from the others. Adding a new field is an easy task now. We just need to add a new method in the builder class.

class BitMaskField {
private:
int data = 0;

public:
void setField(int position){
this->data |= (1 << position);
}

bool isFieldUsed(int position){
int mask = (1 << position);
return (this->data & mask) == mask;
}
};

class BitMaskFieldBuilder {
private:
BitMaskField bitMaskField = BitMaskField();

public:
BitMaskFieldBuilder withField1(){
if(condition1()){
bitMaskField.setField(1);
}
return *this;
}

BitMaskFieldBuilder withField2(){
if(condition2()){
bitMaskField.setField(2);
}
return *this;
}

BitMaskField build(){
return bitMaskField;
}
};

BitMaskField createFields(){
BitMaskField fields = BitMaskFieldBuilder().withField1()
.withField2()
.build();
return fields;
}

Conclusion

In this post, I demonstrated how combining the Bit Mask technique with the Builder design pattern can reduce cyclomatic complexity and simplify the codebase. The benefits are multiples: developers can easily test new implementations without needing to know the details of other fields, the company saves on storage and network transport costs, and users experience faster system responses.

The bit mask field is a valuable technique for minimizing memory and storage usage. However, it can introduce complexity if not implemented thoughtfully. Here, I showed how leveraging abstraction and the Builder design pattern can encapsulate this technique, making it more manageable and reducing overall complexity.

The problem

Creating the bit mask field

Cyclomatic complexity

Builder design pattern

Conclusion

References

Similar Posts